Continuing this week's theme of doing important things without leaving town, I present our results foran exciting project I started earlier this year.

For seven weeks, my coworker Mark Haye and I voluntarily led a class of students here in Tucson, Arizona in an after-school pilot project to teach the ["C" programming language] using [LEGO® Mindstorms® NXT robots]. The ten students, boys and girls ages 9 to 14 years old, were already part of the FIRST [For Inspiration and Recognition of Science and Technology] program, and participated in FIRST Lego League[FLL] robot competitions.Since the students were already familiar building robots, and programming them with a simple graphical system of connecting blocks that perform actions. However, to compete in the next level of robot competitions, FIRST Tech Challenge [FTC],we need to leave this simple graphical programming behind, and upgrade to more precise "C" programming.

Mark is a software engineer for IBM Tivoli Storage Manager and has participated in FLL competitions over the past nine years. This week, he celebrates his 25th anniversary at IBM, and I celebrate my 23rd. The teacher, Ms. Ackerman, and the students referred to us as "Coach Mark" and "Coach Tony".

This was the first time I had worked with LEGO NXT robots. For those not familiar with these robots, you can purchase a kit at your localtoy store. In addition to regular LEGO bricks, beams, and plates, there are motors, wheels, and sensors. A programmable NXT brick has three outputs (marked A,B, and C) to control three motors, and four inputs (marked 1,2,3,4) to receive values from sensors. Programs are written and compiled on laptops and then downloaded to the NXT programmable brick through an USB cable, or wirelessly via Bluetooth.

In the picture shown, an image of the Mars planetary surface is divided into a grid with thick black lines.A light sensor between the front two wheels of the robot is over the black line.

We used the [RobotC programming firmware] and integrated development environment (IDE) from [Carnegie Mellon University].The idea of this pilot was to see how well the students could learn "C". With only a few hours after class on each Wednesday, could we teach young students "C" programming in just seven weeks?

My contribution? I have taught both high school and college classes, and spent over 15 years programming for IBM, so Mark asked me to help.We started with a basic lesson plan:

Week 1	A brief history of the "C" language Understanding statements and syntax Setting motor speed and direction Compiling and downloading your first program
Week 2	Understanding the "while" loop Retrieving input sensor values
Week 3	Understanding the "if-then-else" statement Defining variables with different data types Manipulating string variables Writing a program for the robot to track along a black line on a white background.
Week 4	Understanding local versus global scope variables Writing a program for a robot to count black lines as it crosses them.
Week 5	Defining functions Perform left turns, right turns, and to cross a specific number of lines on a grid pattern to move the robot to a specific location.
Weeks 6 and 7	Mission Impossible: come up with a challenge to make the robot do something that would be difficult to accomplish using the previous NXT visual programming language.

At the completion of these seven weeks, I sat down to interview "Coach Mark"on his thoughts on this pilot project.

Why "C"?

This is a practical programming skill. The "C" language is used throughout the world to program everything from embedded systems to operating systems, and even storage software. This would allow the robots to handle more precise movements, more accurate turns, and more complicated missions.

Can kids learn "C" in only seven weeks?

Part of the pilot project was to see how well the students could understand the material. They were already familiar with building the robots, and understood the basics of programming sensors and motors, so we were hoping this was a good foundation to work from. Some kids managed very well, others struggled.

Did everything go according to plan?

The first two weeks went well, turning on motors and having robots move forward and backward were easy enough. We seemed to lose a few students on week 3, and things got worse from there. However, several of the students truly surprised us and managed to implement very complicated missions. We were quite pleased with the results.

What kind of problems did the kids encounter?

Touch sensor required loops waiting for pressing. Motors did not necessarily turn as expected until more advanced methods were used. Making 90 degree left and right turns accurately was more difficult than expected.

Any funny surprises?

Yes, we had a Challenge Map representing the Mars planetary surface from a previous FLL competition that was dark red and divided into squares with thick black lines. An active light sensor returns a value of "0" (complete darkness) to "100" (bright white).However, the Mars surface had craters that were dark enough to be misinterpreted as a black line causing some unusual results. This required some enhanced programming techniques to resolve.

Did robots help or hurt the teaching process?

I think they helped. Rather than writing programs that just display "Hello World!" on a computer screen, the students can actually see robots move, and either do what they expect, or not!

And when the robots didn't do what they were expected to?

The students got into "debug" mode. They were already used to doing this from previous FLL competitions, but with RobotC, you can leave the USB cable connected (or use wireless Bluetooth) and actually gather debugging information while the robot is running, to see the value of sensors and other variables and help determine why things are not working properly.

Any applicability to the real world of storage?

We have robots in the IBM System Storage TS3500 tape library. These robots scan bar code labels, pull tapes out of shelves and mount them into drives.The programming skills are the same needed for storage software, suchas IBM Tivoli Storage Manager or IBM Tivoli Storage Productivity Center.

The world is becoming smarter, instrumented with sensors, interconnected over a common network, and intelligent enough to react and respond correctly. The lessons of reading sensor values and moving motors can be considered the first step in solutions that help to make a smarter planet.

technorati tags: IBM, C programming language, LEGO, Mindstorms, NXT, FLL, FIRST, robot, RobotC, Bluetooth, USB, TSM, Productivity Center, TS3500, smarter planet

Tags:  infrastructure tape